Andy said:
> My guess is that if it is valuable to have this kind of information in a
> human-readable document, then it is going to be just as valuable to have
> it in a machine-readable schema?
I agree.
I think the DCMI registry will be only one potential "consumer" of the DC
RDF schemas - other applications will retrieve the RDF schemas from a URL
(which may be the namespace name/URI or may be another location which is
obtained from some other data at the namespace URI - how the app finds the
schema isn't the important point here) without ever going near the DCMI
registry.
So this metadata provides some useful information for such an app to use
e.g. an app constructing a data entry form has the information available to
construct a header which identifies the element set. You can perhaps argue
that the app would have that "knowledge" up front otherwise it wouldn't be
going looking for a RDF schema at http://purl.org/dc/elements/1.1/, but I
don't think it's safe to assume that's always going to be the case.
I think we should try to make all our resources (including these RDF
schemas) as "self-describing" as possible - to programs as well as human
readers.
> If the information we want to provide about the schema is really along the
> lines of significant dates, who created it, the name of the schema, etc.
> then why don't we use dc:date, dc:creator and dc:title for goodness sake -
> i.e. use DC metadata to describe the schema!?
But this is exactly what _is_ being provided, isn't it? Contained within the
eor:Schema element are dc: properties i.e.
<rdf:RDF....>
<eor:Schema rdf:about="http://purl.org/dc/elements/1.1/">
<dc:title>The Dublin Core Metadata Element Set</dc:title>
</eor:Schema>
</rdf:RDF>
is just a shorthand (defined in RDF M&S so understood by RDF apps) for
<rdf:RDF....>
<rdf:Description rdf:about="http://purl.org/dc/elements/1.1/">
<rdf:type rdf:resource="......./Schema"/>
<dc:title>The Dublin Core Metadata Element Set</dc:title>
</rdf:Description>
</rdf:RDF>
so an RDF app that doesn't "speak eor" can still obtain _some_ statements
it might "understand" if it "speaks dc". It can ignore the type = Schema
statement.
A more interesting question, maybe (and a variant of one which Andy asked me
this morning!), is: if these are statements about a resource identified by
the URI http://purl.org/dc/elements/1.1/, what exactly is that resource we
are describing here? The namespace? The Dublin Core metadata element set?
The RDF schema representation of the DCMES? A document which is "at the
namespace URI"? I'm not sure....!
Pete
|